机器人 / 具身智能 - arXivDaily 专题

2602.19313 2026-07-24 cs.RO cs.AI cs.LG 版本更新 85%

TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics

TOPReward: 令牌概率作为机器人学中的隐式零样本奖励

Shirui Chen, Cole Harrison, Ying-Chun Lee, Angela Jin Yang, Zhongzheng Ren, Lillian J. Ratliff, Jiafei Duan, Dieter Fox, Ranjay Krishna

机构 * University of Washington（华盛顿大学）； Allen Institute for AI（人工智能研究所）； Amazon（亚马逊）； University of North Carolina at Chapel Hill（北卡罗来纳大学教堂山分校）

专题命中模仿学习与强化学习：robotics(title)；robot learning(abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结 TOPReward通过利用预训练视频视觉-语言模型的令牌概率，提供高效的零样本奖励估计，显著提升机器人任务进度评估的性能和泛化能力。

2604.08958 2026-06-12 cs.LG cs.AI cs.RO 版本更新 85%

WOMBET: World Model-Based Experience Transfer for Robust and Sample-efficient Reinforcement Learning

WOMBET：基于世界模型的经验迁移实现鲁棒且样本高效的强化学习

Mintae Kim, Koushil Sreenath

机构 * Hybrid Robotics, UC Berkeley（混合机器人技术，伯克利大学）

专题命中模仿学习与强化学习：world model(title,abstract)；robotics(abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结提出WOMBET框架，通过源任务中学习世界模型并生成不确定性惩罚的离线数据，再结合自适应采样进行在线微调，实现鲁棒且样本高效的强化学习迁移。

Comments 13 pages, 6 figures, 8th Annual Learning for Dynamics & Control Conference (L4DC)

2604.13733 2026-06-11 cs.LG cs.AI cs.RO 版本更新 85%

Vision-Language-Action Jump-Starting for Reinforcement Learning Robotic Agents

视觉-语言-动作跳跃启动用于强化学习机器人智能体

Angelo Moroncelli, Roberto Zanetti, Marco Maccarini, Loris Roveda

机构 * University of Applied Science and Arts of Southern Switzerland, Department of Innovative Technologies（瑞士南方应用科学与艺术大学创新技术系）； Università della Svizzera Italiana, Faculty of Informatics, Lugano, Switzerland（瑞士意大利大学信息学院，卢加诺，瑞士）

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结提出VLAJS方法，通过稀疏的VLA高层动作建议引导PPO探索，结合方向性动作一致性正则化，提升强化学习在长时域操作任务中的样本效率，并在仿真和真实机器人上验证。

Comments ICRA 2026 Workshop on Reinforcement Learning in the Era of Imitation Learning

2603.16673 2026-05-29 cs.RO cs.AI cs.LG 85%

When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making

机器人何时应该思考？基于强化学习的资源感知推理在具身机器人决策中的应用

Jun Liu, Pu Zhao, Zhenglun Kong, Xuan Shen, Peiyan Dong, Fan Yang, Lin Cui, Hao Tang, Geng Yuan, Wei Niu, Wenbin Zhang, Xue Lin, Gaowen Liu, Yanzhi Wang, Dong Huang

机构 * Robotics Institute, Carnegie Mellon University（卡内基梅隆大学机器人研究所）； Northeastern University（东北大学）； Harvard University（哈佛大学）； Cornell University（康奈尔大学）； MIT（麻省理工学院）； Fujitsu Research of America（美国富士通研究）； Tsinghua University（清华大学）； Peking University（北京大学）； University of Georgia（佐治亚大学）； Florida International University（佛罗里达国际大学）； EmbodyX Inc（EmbodyX公司）； Cisco Systems（思科系统）

专题命中模仿学习与强化学习：robotic(title,abstract)；embodied agent(abstract)；分类 cs.RO、cs.AI、cs.LG

AI总结提出RARRL框架，通过强化学习学习高层编排策略，使具身代理能自适应决定是否调用LLM推理、选择推理角色及分配计算预算，以平衡推理开销与任务成功率。

2510.19732 2025-12-01 cs.AI cs.CV cs.RO 85%

Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning

备忘录：通过强化学习训练内存高效的具身智能体

Gunshi Gupta, Karmesh Yadav, Zsolt Kira, Yarin Gal, Rahaf Aljundi

机构 * University of Oxford（牛津大学）； Georgia Tech University（佐治亚理工学院）； Toyota Motor Europe（丰田欧洲公司）

专题命中模仿学习与强化学习：embodied agent(title,abstract)；navigation(abstract)；分类 cs.RO、cs.AI、cs.CV

AI总结 Memo通过在训练过程中交错周期性总结标记与输入，实现内存高效的具身智能体强化学习训练，优于长上下文基线并更高效。

Comments Accepted for Spotlight Presentation at NeurIPS 2025

2412.20338 2024-12-31 cs.RO cs.AI cs.LG 85%

Exploiting Hybrid Policy in Reinforcement Learning for Interpretable Temporal Logic Manipulation

Hao Zhang, Hao Wang, Xiucai Huang, Wenrui Chen, Zhen Kan

专题命中模仿学习与强化学习：manipulation(title,abstract)；robot learning(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Accepted by IROS 2024. Code:https://github.com/Charlie0257/HyTL

2412.09858 2024-12-16 cs.RO cs.AI cs.LG 85%

RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning

Charles Xu, Qiyang Li, Jianlan Luo, Sergey Levine

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

2406.12499 2024-06-19 cs.LG cs.AI cs.RO 85%

Autonomous navigation of catheters and guidewires in mechanical thrombectomy using inverse reinforcement learning

Harry Robertshaw, Lennart Karstensen, Benjamin Jackson, Alejandro Granados, Thomas C. Booth

专题命中模仿学习与强化学习：navigation(title,abstract)；robotics(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Abstract shortened for arXiv character limit

Journal ref Int J CARS (2024)

2403.12891 2024-03-20 cs.RO cs.AI cs.CV 85%

Adaptive Visual Imitation Learning for Robotic Assisted Feeding Across Varied Bowl Configurations and Food Types

Rui Liu, Amisha Bhaskar, Pratap Tokekar

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.CV

2303.06710 2023-03-15 cs.RO cs.AI cs.LG 85%

Decision Making for Human-in-the-loop Robotic Agents via Uncertainty-Aware Reinforcement Learning

Siddharth Singi, Zhanpeng He, Alvin Pan, Sandip Patel, Gunnar A. Sigurdsson, Robinson Piramuthu, Shuran Song, Matei Ciocarlie

专题命中模仿学习与强化学习：robotic(title,abstract)；navigation(abstract)；分类 cs.RO、cs.AI、cs.LG

2212.08244 2022-12-19 cs.RO cs.CV cs.LG 85%

Offline Reinforcement Learning for Visual Navigation

Dhruv Shah, Arjun Bhorkar, Hrish Leen, Ilya Kostrikov, Nick Rhinehart, Sergey Levine

专题命中模仿学习与强化学习：navigation(title,abstract)；robotic(abstract)；分类 cs.RO、cs.CV、cs.LG

Comments Project page https://sites.google.com/view/revind/home

2210.08217 2022-11-28 cs.RO cs.AI cs.IT cs.LG math.IT 85%

PI-QT-Opt: Predictive Information Improves Multi-Task Robotic Reinforcement Learning at Scale

Kuang-Huei Lee, Ted Xiao, Adrian Li, Paul Wohlhart, Ian Fischer, Yao Lu

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments CoRL 2022. 21 pages, 9 figures. The supplementary video is available at https://kuanghuei.github.io/piqtopt

2207.08974 2022-07-20 cs.HC 85%

ARtonomous: Introducing Middle School Students to Reinforcement Learning Through Virtual Robotics

Griffin Dietz, Jennifer King Chen, Jazbo Beason, Matthew Tarrow, Adriana Hilliard, R. Benjamin Shapiro

专题命中模仿学习与强化学习：robotics(title,abstract)；navigation(abstract)；robotic(abstract)

Comments In Proceedings of Interaction Design and Children (IDC '22)

2109.09180 2022-04-07 cs.LG cs.AI cs.RO 85%

Lifelong Robotic Reinforcement Learning by Retaining Experiences

Annie Xie, Chelsea Finn

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Supplementary website at https://sites.google.com/view/retain-experience/

2105.00931 2021-10-14 cs.CV cs.AI cs.LG cs.MA 85%

GridToPix: Training Embodied Agents with Minimal Supervision

Unnat Jain, Iou-Jen Liu, Svetlana Lazebnik, Aniruddha Kembhavi, Luca Weihs, Alexander Schwing

专题命中模仿学习与强化学习：embodied agent(title)；embodied AI(abstract)；navigation(abstract)；分类 cs.AI、cs.CV、cs.LG

Comments Project page: https://unnat.github.io/gridtopix/ ; last two authors contributed equally

2011.05782 2020-11-12 cs.RO cs.AI cs.LG 85%

Reinforcement Learning Experiments and Benchmark for Solving Robotic Reaching Tasks

Pierre Aumjaud, David McAuliffe, Francisco Javier Rodríguez Lera, Philip Cardiff

专题命中模仿学习与强化学习：robotic(title,abstract)；robotics(abstract)；分类 cs.RO、cs.AI、cs.LG

Journal ref Advances in Intelligent Systems and Computing, 1285 (2021), 318-331

1909.09282 2019-09-23 cs.RO cs.AI cs.LG cs.SY eess.SY 85%

How Much Do Unstated Problem Constraints Limit Deep Robotic Reinforcement Learning?

W. Cannon Lewis, Mark Moll, Lydia E. Kavraki

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments Rice University technical report

1905.07193 2019-05-20 cs.LG cs.AI cs.RO stat.ML 85%

MaMiC: Macro and Micro Curriculum for Robotic Reinforcement Learning

Manan Tomar, Akhil Sathuluri, Balaraman Ravindran

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments To appear in the Proceedings of the 18th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2019). (Extended Abstract)

1904.07854 2019-05-17 cs.LG cs.CV cs.RO stat.ML 85%

End-to-End Robotic Reinforcement Learning without Reward Engineering

Avi Singh, Larry Yang, Kristian Hartikainen, Chelsea Finn, Sergey Levine

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.CV、cs.LG

Comments Accepted to RSS 2019. 14 pages and 13 figures including references and appendix. Website: https://sites.google.com/view/reward-learning-rl/home

1811.10092 2019-04-09 cs.CV cs.AI cs.CL cs.RO 85%

Reinforced Cross-Modal Matching and Self-Supervised Imitation Learning for Vision-Language Navigation

Xin Wang, Qiuyuan Huang, Asli Celikyilmaz, Jianfeng Gao, Dinghan Shen, Yuan-Fang Wang, William Yang Wang, Lei Zhang

专题命中模仿学习与强化学习：navigation(title,abstract)；embodied agent(abstract)；分类 cs.RO、cs.AI、cs.CV

Comments CVPR 2019 Oral

1702.06329 2017-11-08 cs.AI cs.LG cs.RO 85%

Towards a Common Implementation of Reinforcement Learning for Multiple Robotic Tasks

Angel Martínez-Tenor, Juan Antonio Fernández-Madrigal, Ana Cruz-Martín, Javier González-Jiménez

专题命中模仿学习与强化学习：robotic(title,abstract)；robotics(abstract)；分类 cs.RO、cs.AI、cs.LG

Comments 15 pages, 10 figures, 7 tables. To be published in a scientific journal

2410.13816 2025-02-26 cs.RO cs.LG 84%

Steering Your Generalists: Improving Robotic Foundation Models via Value Guidance

Mitsuhiko Nakamoto, Oier Mees, Aviral Kumar, Sergey Levine

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.LG；robot learning(comments)

Comments Conference on Robot Learning (CoRL) 2024. Project Page: https://nakamotoo.github.io/V-GPS

2408.03539 2024-09-17 cs.RO cs.LG 84%

Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes

Chen Tang, Ben Abbatematteo, Jiaheng Hu, Rohan Chandra, Roberto Martín-Martín, Peter Stone

专题命中模仿学习与强化学习：robotics(title,abstract)；robotic(abstract)；分类 cs.RO、cs.LG

Comments The first three authors contributed equally. Accepted to Annual Review of Control, Robotics, and Autonomous Systems

2109.11178 2021-11-08 cs.RO cs.LG 84%

Hierarchies of Planning and Reinforcement Learning for Robot Navigation

Jan Wöhlke, Felix Schmitt, Herke van Hoof

专题命中模仿学习与强化学习：navigation(title,abstract)；robotic(abstract)；分类 cs.RO、cs.LG；robotics(comments)

Comments 7 pages, 5 figures, 2021 IEEE International Conference on Robotics and Automation (ICRA), v2: DOI number added

2002.05630 2021-02-01 cs.AI cs.LG stat.ML 84%

On the Sensory Commutativity of Action Sequences for Embodied Agents

Hugo Caselles-Dupré, Michael Garcia-Ortiz, David Filliat

专题命中模仿学习与强化学习：embodied agent(title,abstract)；robotic(abstract)；分类 cs.AI、cs.LG；robot learning(comments)

Comments Accepted to RSS'20 Workshop on Self-Supervised Robot Learning & to the Workshop on Learning in Artificial Open Worlds at ICML20 & Extended abstract at AAMAS21

1803.07635 2018-07-26 cs.RO cs.LG 84%

Learning Robotic Assembly from CAD

Garrett Thomas, Melissa Chien, Aviv Tamar, Juan Aparicio Ojea, Pieter Abbeel

专题命中模仿学习与强化学习：robotic(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.LG；robotics(comments)

Comments In the proceedings of the IEEE International Conference on Robotics and Automation (ICRA), Brisbane, Australia, May 2018

2607.02431 2026-07-03 cs.RO cs.AI 新提交 84%

WorldSample: Closed-loop Real-robot RL with World Modelling

WorldSample：基于世界建模的闭环真实机器人强化学习

Yuquan Xue, Le Xu, Zeyi Liu, Zhenyu Wu, Zhengyi Gu, Xinyang Song, Bofang Jia, Ziwei Wang

机构 * PINE Lab, School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore（南洋理工大学电气与电子工程学院PINE实验室）； Department of Electronic Engineering, Tsinghua University, Beijing, China（清华大学电子工程系）； School of Automation, Central South University, Changsha, China（中南大学自动化学院）； School of Automation, Beijing University of Posts and Telecommunications, Beijing, China（北京邮电大学自动化学院）

专题命中模仿学习与强化学习：world model(title,abstract)；manipulation(abstract)；分类 cs.RO、cs.AI

AI总结提出WorldSample框架，通过物理 rollout、世界模型生成与策略改进的闭环，结合策略节奏学习（PPL）调节训练，在接触性精密操作任务中成功率提升28%，训练步数减少59%。

Comments 16 pages, 9 figures, conference paper

2602.13977 2026-06-30 cs.RO cs.AI 84%

WoVR: World Models as Reliable Simulators for Post-Training VLA Policies with RL

WoVR：基于世界模型的可靠模拟器用于训练后VLA策略的强化学习

Zhennan Jiang, Shangqing Zhou, Yutong Jiang, Zefang Huang, Mingjie Wei, Yuhui Chen, Tianxing Zhou, Zhen Guo, Hao Lin, Quanlu Zhang, Yu Wang, Haoran Li, Chao Yu, Dongbin Zhao

专题命中模仿学习与强化学习：world model(title,abstract)；robotic(abstract)；分类 cs.RO、cs.AI

AI总结本文提出WoVR框架，通过可控动作条件视频世界模型和关键帧初始化回放提升模拟稳定性，实现稳定长周期模拟回放和有效策略优化，取得优于LIBERO的性能。

Comments 25pages, 11 figures

2511.15279 2026-04-06 cs.RO cs.CV 84%

Look, Zoom, Understand: The Robotic Eyeball for Embodied Perception

看、聚焦、理解：用于具身感知的机器人眼球

Jiashu Yang, Yifan Han, Yucheng Xie, Ning Guo, Wenzhao Lian

机构 * School of Artificial Intelligence, Shanghai Jiao Tong University（上海交通大学人工智能学院）； Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）； Dalian University of Technology（大连理工大学）

专题命中模仿学习与强化学习：robotic(title,abstract)；embodied AI(abstract)；分类 cs.RO、cs.CV

AI总结本文提出EyeVLA框架，通过整合视觉感知、语言理解和物理摄像头控制，实现语言引导的主动视觉感知。该框架在500个真实样本上训练，使机器人在50种不同场景中完成任务的平均完成率为96%。

2603.20607 2026-03-24 cs.RO cs.LG 84%

Towards Practical World Model-based Reinforcement Learning for Vision-Language-Action Models

面向视觉-语言-动作模型的实用世界模型强化学习

Zhilong Zhang, Haoxiang Ren, Yihao Sun, Yifei Sheng, Haonan Wang, Haoxin Lin, Zhichao Wu, Pierre-Luc Bacon, Yang Yu

机构 * National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China（新型软件技术国家实验室，南京大学，南京，中国）； School of Artificial Intelligence, Nanjing University, Nanjing, China（人工智能学院，南京大学，南京，中国）； Mila - Quebec AI Institute（魁北克AI研究所）

专题命中模仿学习与强化学习：world model(title,abstract)；robotic(abstract)；分类 cs.RO、cs.LG

AI总结本文提出VLA-MBPO框架，解决视觉-语言-动作模型在强化学习中的世界建模、多视角一致性及稀疏奖励下的误差累积问题，提升策略性能和样本效率。

1. 模仿学习与强化学习 4106 篇

TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics

WOMBET: World Model-Based Experience Transfer for Robust and Sample-efficient Reinforcement Learning

Vision-Language-Action Jump-Starting for Reinforcement Learning Robotic Agents

When Should a Robot Think? Resource-Aware Reasoning via Reinforcement Learning for Embodied Robotic Decision-Making

Memo: Training Memory-Efficient Embodied Agents with Reinforcement Learning

Exploiting Hybrid Policy in Reinforcement Learning for Interpretable Temporal Logic Manipulation

RLDG: Robotic Generalist Policy Distillation via Reinforcement Learning

Autonomous navigation of catheters and guidewires in mechanical thrombectomy using inverse reinforcement learning

Adaptive Visual Imitation Learning for Robotic Assisted Feeding Across Varied Bowl Configurations and Food Types

Decision Making for Human-in-the-loop Robotic Agents via Uncertainty-Aware Reinforcement Learning

Offline Reinforcement Learning for Visual Navigation

PI-QT-Opt: Predictive Information Improves Multi-Task Robotic Reinforcement Learning at Scale

ARtonomous: Introducing Middle School Students to Reinforcement Learning Through Virtual Robotics

Lifelong Robotic Reinforcement Learning by Retaining Experiences

GridToPix: Training Embodied Agents with Minimal Supervision

Reinforcement Learning Experiments and Benchmark for Solving Robotic Reaching Tasks

How Much Do Unstated Problem Constraints Limit Deep Robotic Reinforcement Learning?

MaMiC: Macro and Micro Curriculum for Robotic Reinforcement Learning

End-to-End Robotic Reinforcement Learning without Reward Engineering

Reinforced Cross-Modal Matching and Self-Supervised Imitation Learning for Vision-Language Navigation

Towards a Common Implementation of Reinforcement Learning for Multiple Robotic Tasks

Steering Your Generalists: Improving Robotic Foundation Models via Value Guidance

Deep Reinforcement Learning for Robotics: A Survey of Real-World Successes

Hierarchies of Planning and Reinforcement Learning for Robot Navigation

On the Sensory Commutativity of Action Sequences for Embodied Agents

Learning Robotic Assembly from CAD

WorldSample: Closed-loop Real-robot RL with World Modelling

WoVR: World Models as Reliable Simulators for Post-Training VLA Policies with RL

Look, Zoom, Understand: The Robotic Eyeball for Embodied Perception

Towards Practical World Model-based Reinforcement Learning for Vision-Language-Action Models